Final Report Summary - RNA+P=123D (Breaking the code of RNA sequence-structure-function relationships: New strategies and tools for modelling and engineering of RNA and RNA-protein complexes)
The main goal of the project “RNA+P=123D: Breaking the code of RNA sequence-structure-function relationships: New strategies and tools for modelling and engineering of RNA and RNA-protein complexes”, was to develop and validate a toolkit of methods for studying structure of RNA and RNA-protein complexes. The plans for the development of the toolkit were divided into three Principal Aims: 1) The development of theoretical methods for prediction and design of the three dimensional structures of RNA molecules, free and in complexes with proteins; 2) the implementation of experimental methods that provide data that can be used as restraints in the modeling process or to validate theoretical predictions; 3) the development of a computational workflow for testing the predictions. The project has achieved all three Principal Aims.
Efforts were mostly concentrated on P.A.1 and this work has led to the development of an extensive set of new computational methods. The main technological achievement of P.A.1. was the development of a complete toolbox of methods for modeling and design of RNA 3D structures, and for modeling RNA complexes with other molecules, in particular large complexes including proteins. The tools developed within this project have different purpose, from a physical potential for high-resolution refinement, to low-resolution methods that operate on individual domains as rigid bodies. In particular, we finalized the development of a completely new version of the SimRNA method for RNA folding, and made it publicly available both as a standalone tool, and as a web server. We also developed a prototype of an RNA template-free modeling method called RNA Masonry, based on the fragment assembly approach, which uses motifs from the RNA Bricks database as building blocks and the SimRNA potential as a scoring method of the assembled structures. We developed and made available the QRNAS method, which optimizes models of RNA structure (e.g. those built with SimRNA or RNA Masonry) using a high-resolution physical potential. We developed the NPDock server for protein-nucleic acid docking that treats molecules as rigid bodies, as well as a method for protein-RNA complex modeling derived from SimRNA, called SimRNP, which enables flexible modeling of both RNA and protein molecules as well as complexes comprising both types of molecules. We have also developed a method, which designs RNA sequences that maximize the probability of folding into a user-defined 3D structure. This method includes two components: DesiRNA for the design on the secondary structure level and an extension of SimRNA called SimRNA-design, which allows mutations of RNA sequence in the course of a 3D structure folding simulation. Also a number of other accessory software tools for analyzing RNA structures were developed.
The experimental validation of methods developed in the course of this project (P.A.2) included multiple projects, in which experiments we carried out, as planned – partially in our laboratory, and partially by collaborators. We implemented in the laboratory a number of biochemical and biophysical methods for characterization of RNA and RNA-protein complexes, including chemical probing methods as well as RNA and RNP crystallography. Experimental analyses involved a number of RNA-protein complexes with vastly different sizes. The biggest successes achieved involved the characterization of RNases (MiniIII and RNazeH-zinc finger hybrid) that act sequence-specifically on RNA, and can be used for targeted RNA cleavage essentially like restriction enzymes are used for DNA cleavage. Successful modeling of large macromolecular complexes based on experimental data is exemplified by the CCR4-Not complex.
The RNA+P=123D project has also involved the development of methods for benchmarking of methods for RNA 3D structure prediction. In particular, we developed a workflow called CompaRNA, for continuous assessment of RNA secondary structure prediction methods, and implemented it as a web server. We have also participated in the benchmarking of RNA 3D structure prediction methods in the RNA Puzzles experiment, which has been recently expanded to include web servers.
Efforts were mostly concentrated on P.A.1 and this work has led to the development of an extensive set of new computational methods. The main technological achievement of P.A.1. was the development of a complete toolbox of methods for modeling and design of RNA 3D structures, and for modeling RNA complexes with other molecules, in particular large complexes including proteins. The tools developed within this project have different purpose, from a physical potential for high-resolution refinement, to low-resolution methods that operate on individual domains as rigid bodies. In particular, we finalized the development of a completely new version of the SimRNA method for RNA folding, and made it publicly available both as a standalone tool, and as a web server. We also developed a prototype of an RNA template-free modeling method called RNA Masonry, based on the fragment assembly approach, which uses motifs from the RNA Bricks database as building blocks and the SimRNA potential as a scoring method of the assembled structures. We developed and made available the QRNAS method, which optimizes models of RNA structure (e.g. those built with SimRNA or RNA Masonry) using a high-resolution physical potential. We developed the NPDock server for protein-nucleic acid docking that treats molecules as rigid bodies, as well as a method for protein-RNA complex modeling derived from SimRNA, called SimRNP, which enables flexible modeling of both RNA and protein molecules as well as complexes comprising both types of molecules. We have also developed a method, which designs RNA sequences that maximize the probability of folding into a user-defined 3D structure. This method includes two components: DesiRNA for the design on the secondary structure level and an extension of SimRNA called SimRNA-design, which allows mutations of RNA sequence in the course of a 3D structure folding simulation. Also a number of other accessory software tools for analyzing RNA structures were developed.
The experimental validation of methods developed in the course of this project (P.A.2) included multiple projects, in which experiments we carried out, as planned – partially in our laboratory, and partially by collaborators. We implemented in the laboratory a number of biochemical and biophysical methods for characterization of RNA and RNA-protein complexes, including chemical probing methods as well as RNA and RNP crystallography. Experimental analyses involved a number of RNA-protein complexes with vastly different sizes. The biggest successes achieved involved the characterization of RNases (MiniIII and RNazeH-zinc finger hybrid) that act sequence-specifically on RNA, and can be used for targeted RNA cleavage essentially like restriction enzymes are used for DNA cleavage. Successful modeling of large macromolecular complexes based on experimental data is exemplified by the CCR4-Not complex.
The RNA+P=123D project has also involved the development of methods for benchmarking of methods for RNA 3D structure prediction. In particular, we developed a workflow called CompaRNA, for continuous assessment of RNA secondary structure prediction methods, and implemented it as a web server. We have also participated in the benchmarking of RNA 3D structure prediction methods in the RNA Puzzles experiment, which has been recently expanded to include web servers.